NUREG/IA-00 18

InteionationalAgreement Report

RELAP5/MOD2 Assessment,OECD-LOFT Small BreakExperiment LP-SB-03

Prepared byS. Guntay

Paul Scherrer InstituteCH-5303 Wurenlingen, Switzerland

Office of Nuclear Regulatory ResearchU.S. Nuclear Regulatory CommissionWashington, DC 20555

April 1990

Prepared as part ofThe Agreement on Research Participation and Technical Exchangeunder the International Thermai-Hydraulic Code Assessmentand Application Program (ICAP)

Published byU.S. Nuclear Regulatory Commission

M

NOTICE

This report was prepared under an international cooperativeagreement for the exchange of technical information. Neitherthe United States Government nor any agency thereof, or any oftheir employees, makes any' warranty, expressed or implied, orassumes any legal liability or responsibility for any third party'suse, or the results of such use, of any information, apparatus pro-duct or process disclosed in this report, or represents that its useby such third party Would not infringe privately owned rights.

I

Available from

Superintendent of DocumentsU.S. Government Printing Office

P.O. Box 37082Washington, D.C. 20013-7082

and

National Technical Information ServiceSpringfield, VA 22161

NUREG/IA-001 8

InternallonalAgreement Report

RELAP5/MOD2 Assessment,OECD-LOFT Small BreakExperiment LP-SB-03Prepared byS. Guntay

Paul Scherrer InstituteCH-5303 Wurenlingen, Switzerland

Office of Nuclear Regulatory ResearchU.S. Nuclear Regulatory CommissionWashington, DC 20555

April 1990

Prepared as part ofThe Agreement on Research Participation and Technical Exchangeunder the International Thermal-Hydraulic Code Assessmentand Application Program (ICAP)

Published byU.S. Nuclear Regulatory Commission

NOTICE

This report is based on work performed under the sponsorship of the

Swiss Federal Office of Energy. The information in this report has

been provided to the USNRC under the terms of the International

Code Assessment and Application Program (ICAP) between the United

States and Switzerland (Research Participation and Technical

Exchange between the United States Nuclear Regulatory Commission

and the Swiss Federal Office of Energy in the field of reactorsafety research and development, May 1985). Switzerland has

consented to the publication of this report as a USNRC document in

order to allow the widest possible circulation among the reactor

safety community. Neither the United States Government nor

Switzerland or any agency thereof, or any of their employees, makesany warranty, expressed or implied, or assumes any legal liability

of responsibility for any third party's use, or the results of suchuse, or any information, apparatus, product or process disclosed

in this report, or represents that its use by such third party

would not infringe privately owned rights.

RELAP5/MOD2 Assessment. OECD-LOFT Small Break Experiment

LP-SB-3

S. G~intay

ABSTRACT

An analysis of the experimental results and oost-test calculationsusing RELAP5/MOD2 carried out for OECD-LOFT small break experimentLP-SB-3 are presented. Experiment LP-SB-3 was conducted onMarch 5, 1984 in the.Loss-of -Fluid Test (LOFT) facility located at theIdaho National Engineering Laboratory (INEL). The experiment simulated asmall cold leg break, with concurrent loss of high pressure injectionsystem, and cooldown and recovery by feed and bleed of the steamgenerator secondary side and accumulator injection, respectively.

The analysis was under taken as a part of a program at EIR aimed atdeveloping experience in using the latest generation of best estimate Lossof Coolant Accident (LOCA) analysis computer codes, and to improveunderstanding of Small Break LOCA transients and as well as a part of aprogram aimed at assessing the RELAP5/MOD2 code. The latest availableversion (Cycle 33 to 36.1) of the code was used. The particular testselected for the analysis included several phenomena potentially relevantto any PWR plant operating in Switzerland.

This report documents a short post-test analysis of the experimentemphasizing the results of additional analysis prerformed during the courseof this task. RELAPS/MOD2 input model and results of the post-test

calculation are documented. Included in the report is the results of asensitivity analysis which show the predicted thermal-hydraulic response

to a different input model.

SUMMARY

This report documents the post-test calculation of OECD-LOFTExperiment LP-SB-3 using the RELAP5/MOD2 computer code.

Experiment LP-SB-3 simulated a hypothetical loss-of-coolant accident(LOCA) which resulted from a 4.67 cm diameter single-ended break in thecold leg of a large commercial pressurized water reactor (PWR) with aconcurrent loss of high pressure emergency core coolant injectioncapability. The experiment was initiated from conditions which werescaled to be representative of the design operating conditions of a largePWR; reactor power of 50 MWe, core temperature rise of 20 K, and systempressure of 14.95 MPa. The experiment provided data on (a) core heattransfer characteristics when core uncovery occurs during relatively slowboil-off conditions, (b) effectiveness of steam generator feed and bleed todepressurize and cool a highly voided system, (c) effectiveness of theaccumulator injection in establishing core cooling in a highly voidedsystem.

Results of Experiment LP-SB-3 showed that decay heat removal fromthe 4.66 cm diameter equivalent small break LOCA was dominated by thebreak flow and heat transfer to the secondary side. It was believed thatduring the boil-off phase, reflux condensation took place for most of theperiod. The possibility of condensed liquid draining back through the hotleg and into the core was supported by the asymmetrical fuel claddingtemperature developments. The feed and bleed of the steam generatorsecondary side was effective depressurizing the system. The accumulatorinjection established core cooling.

The RELAP5/MOD2 computer code was shown to be valuable inunderstanding the physical phenomena in the experiment. Althoughdifferences in detail were observed between the celculational results andexperimental data, the code generally performed well, predicting all the.key events in the correct sequence and with reasonable timing.

ii

TABLE OF CONTENTS

A bstract .................................................................................................................. iSum m ary .................................................................................................................. ii1. Introduction ..................................................................................................... .12. Description and Discussion of Experiment LP-SB-3 3

2.1 Description of Experiment LP-SB-3 ......................................... 32.2 Discussion of the Experimental Results .................................. 4

2.2.1 Mass Depletion Phase ...................................................... 42.2.2 Core Boil-Off Phase ......................................................... 62.2.3 Discussion of Primary-Secondary Relationship

During Experiment LP-SB-3 ......................................... .72.2.3. I Reflux Condensation And Heat Transfer

to the Secondary Side ......................................2.2.3.2 Liquid Drainback and its Effects on Fuel

Rod Temperatures ............................................. 92.2.4 Core Cooldown and Recovery Phase ............................ 9

3. RELAP5/MOD2 Computer Code Simulation of ExperimentLP-SB-3 ........................................................................................................... 1 03.1 RELAP5/MOD2 Description ........................................................ ....... 103.2 RELAP5/MOD2 Nodalization and Input Model for

Experiment LP-SB-3 .................................................................... ....... 114. Results of the RELAP5/MOD2 Post-Test Calculation .................... 12

4.1 Calculation of the Steady State ............................................. ....... 124.2 Transient Calculation ...................................................................... 14

4.2.1 Mass Depletion Phase ...................................................... 154.2.1.1 Primary and Secondary System Pressure

Behavior .................................... ..... 154.2.1.2 Break Flow .......................................................... 164.2.1 .3 Density ................................................................. 164.2.1.4 Pump Behavior ................................................... 174.2.1.5 System Mass Inventory ................................... 17

4.2.2 Core Boil-Off Phase ........................................................ 174.2.2. 1 Primary-Secondary System Pressure

Benavior ........................................................ ...... 174.2.2.2 Density ................................................................. 18

4.2.2.3 Break Flow ...................................................... . .... . 194.2.2.4 Core Dryout and Thermal Response ............. 19

4.2.3 Core Cooling and Recovery Phase .......................... ...... 224.2.3.1 Primary and Secondary System Pressure

Behavior ........................................................... ....... 224.2.3.2 'Core Thermal Behavior ............... ...... 234.2.3.3 Accumulator and LPIS Injection .................. 244.2.3.4 Accumulator and LPIS Flow ........................... 24

5. Sensitivity Calculation ............................................................................ 245. 1 Pressures ............................................................................................... 255.2 Core Uncovery ..................................................................................... 25

6. Run Statistics ............................................................................................... 267. Conclusions and Recommendations ....................................................... 278. References ..................................................................................................... 28

T ables .................................................................................................... . ........ 30Figures ...................................................................................................... . ....... 32Appendix A ............................................................................................. . ........ 84

TABLES

I. Chronology of the events2. Initial conditions

FIGURES

1. Primary and secondary pressures (0-150 s)2. Primary and secondary pressures (0-7000 s)3. Steam generator pressure and.MSCV cycling set points4. Fluid density in the break piping5. Break flow6. Fluid density in the intact loop hot leg7. Fluid density in the intact loop cold leg6. Fluid velocity in the intact loop hot leg9. Pressure rise across the primary coolant pumps10. Fluid density in the intact loop seal11. Fuel cladding temperatures in the center bundle12. Cladding temperatures at different radial positions13. Primary secondary pressure differential14. Temperature difference between the saturation temperature of the

primary system and the temperature of secondary system15. Steam superheat in the steam generator inlet plenum16. Response of reactor vessel liquid level probes, core dryout and quench

data from thermocouples17. RELAP5/MOD2 nodalization for Experiment LP-SB-318.19.20.21.22.23.24.25.26.27.28.

Comparison ofComparison ofComparison ofComparison ofComparison ofComparison of

Comparison ofComparison of

Comparison ofComparison ofComparison ofsystems

pressures in the intact loop hot legpressures in the steam generator secondary sidebreak flowsdensities in the intact loop hot legdensities in the intact loop cold legdensities in the beak pipingpressure differential across pumpsystem mass inventoriespressures in the intact loop hot legpressures in the steam generator secondary sidecalculated pressures in the primary and secondary

V

29. Calculated temperature difference between the primary and secondarysystems

30. Vapor generation rate in the steam generator upright31. Calculated liquid flow rate in the steam generator inlet plenum32. Calculated liquid and vapor flow rates in the intact loop hot leg33. Comparison of densities in the intact loop hot leg34. Comparison of.densities in the intact loop cold leg35. Comparison of densities in the loop seal36. Calculated collapsed liquid level in the reactor vessel37. Comparison of break flow rates38. Comparison of dryouts measured and calculated39. Comparison of fuel cladding tempgratures at 12.7 cm elevation40. Comparison of fuel cladding temperatures at 38 cm elevation41. Comparison of fuel cladding temperatures at 69 cm elevation42. Comparison of fuel cladding temperatures at 102 cm elevation43. Comparison of fuel cladding temperatures at 114 cm elevation44. Comparison of fuel cladding temperatures at 157 cm elevation45. Comparison of pressures in the intact loop hot leg46. Comparison of pressures in the secondary side47. Comparison of ECC injection flow rates48. Comparison of integrated ECC injection flows49. RELAP5/MOD2 nodalization for the sensitivity analysis50. Comparison of calculated fuel cladding temperatures at level 55 1. Comparison of calculated fuel cladding temperatures at level 352. CPU time versus transient time and required CPU seconds per real

secondA-I. LOFT configuration for cold leg intact loop small break LP-SB-3A-2. Experiment break piping instrument configurationA-3. LOFT core configuration and instrumentation

vi

1. INTRODUCTION

The Loss-of-Fluid Test (LOFT) experimental facility is a 50 MW testreactor at the Idaho National Engineering Laboratory (INEL) used for thestudy of Loss of Coolant Accidents, and other reactor transients. The LOFTreactor system, in particular the primary coolant system and reactor core,is a fully operational (scaled) representation of a commercial pressurizedwater reactor (PWR). It has a nuclear powered core and a single active(the normally intact loop), containing a U tube Steam Generator (SG) andtwo reactor coolant pumps in parallel branches in the cold leg. A secondloop (the normally broken loop), containing passive hydraulic resistancesto simulate the steam generator and pumps, is also connected, primarilyfor use in large break LOCA simulations.

LOFT experiment LP-SB-31 simulated a 4.66 cm (1.84 in) break in thecold leg of a commercial PWR. Additional features simulated were thefailure of high pressure emergency core cooling injection system, and feedand bleed of the steam generator secondary side. The experiment wasconducted on March 5, 1984 for the Organization for EconomicCorporation and Development (OECD) consortium. This report reviewsbriefly the post test analysis carried out at INEL together with additionalpost test analysis carried out during this study and the post testcalculations performed using the RELAP5/MOD2 2 (Cycles 33 to 36. 1). Thedata deck used for the analysis was based on pre-test prediction deck usedby the INEL to produce Best Estimate Prediction Document (EPD) 3 . SeveralImprovements and corrections on the pre-test deck were done (a) toconvert the deck to the RELAP5/MOD2 format, (b) to 'eliminate some minorerrors, (c) to tune various operations (such as opening or closure ofvalves) according to the experimental data, (d) to apply cross cross flowjunction model.

The analysis was undertaken to assess the RELAP5/MOD2 computercode, in displaying a wide range of thermal-hydraulic phenomena. Theseincludes:

I

a- Primary-secondary relationship Single and two phase forcedconvectionReflux natural circulation

b-Slow boil-offc-Core heat transferd-Single and two phase break flowe-Pump performance, degradationf-Cooldown and recovery of the plant by steam generator feed and

bleed and accumulator injection

Section 2 of this document presents a short description of ExperimentLP-SB-3 and discusses briefly the post-test analysis of the experimentemphasizing the new findings during the conduction of this study. TheRELAP5/MOD2 code is introduced in Section 3 along with the nodalizationused in the base case calculation. Section 4 discusses the results ofpost-test calculation and compares the results of the calculation with thedata. Section 5 discusses the sensitivity study performed and comparesthe base and sensitivity case results. Section 6 discusses run statistics.Conclusions and recommendations are discussed in Section 7. Adescription of the LOFT facility is presented in Appendix A. Appendix Bcontains details of the input data for the base and sensitivity analyses.

2

2. DESCRIPTION AND DISCUSSION OF EXPERIMENT LP-SB-3

OECD-LOFT experiment series LP-SB-3 was designed to investigate thesystem thermal-hydraulic response under (a) slow coolant boil-off leadingto core uncovery at high system pressure, (b) steam generator feed andbleed operation, and (c) plant recovery with accumulator injection.

A short description of Experiment LP-SB-3 and discussion of post-testanalysis of the experiment beyond what was already presented in theQuick Look Report (QLR)I are presented in the following subsections.

2.1 Description of Experiment LP-SB-3

Experiment LP-SB-3 simulated a small cold leg break loss-of-coolanttransient, with a scaled break size corresponding to a 4.66 cm (1.84 in.)pipe diameter in a reference commercial PWR. The experiment wasinitiated from conditions representative of those in a commercial PWR byopening a valve in the normally intact loop cold leg break piping. The scramwas initiated on low pressure in the intact loop hot leg. The primarysystem depressurized rapidly until fluid saturation conditions werereached in the hot leg at about 100s, which resulted in a decrease in theprimary system depressurization rate. As a result, a reduction of the breakmass flow rate occurred with void formation in the coolant. Since theprimary system energy loss from the break due to the size of the breakwas not enough to remove the decay heat energy, the heat transfer to thesecondary side resulted in a higher pressurization of the secondary sidewhich caused the main steam valve to cycle four times and to relieve theenergy of the system. The primary coolant pumps continued operation untilthe system mass inventory was reduced to approximately 2Q00 Kg. Singlephase forced convection before the system reached saturated conditionsand two phase forced convection until the pump trip were observed.Following the pump trip, reflux condensation occurred depending on theprimary-secondary relationship. The pumps during the two phase systemcoolant condition delivered a relatively homogeneous mixture of liquid and

3

steam to the intact loop cold leg and the piping upstream of the breaknozzle. The primary and secondary system pressures became almost thesame after the primary system reached saturation coolant conditions andremained coupled during the rest of the transient. A top to bottom slowcore uncovery was observed after the system mass inventory was reducedto approximately 1750 Kg. Steam generator feed and bleed operation wasinitiated when the peak cladding temperature reached 987 K which let to arapid cooldown and depressurization of the secondary side followedclosely by the primary side. A partial rewet of the core from belowfollowed the initiation of the feed and bleed. Primary pressure droppedbelow the accumulator set point shortly after this, causing a furtherpressure reduction. The accumulator injection caused core rewet from thebottom upwards and quenched the core quickly to the coolant saturationtemperature. Low pressure emergency core cooling injection system(LPIS) became operational after the complete core quench at a systempressure of 1.03 MPa and provided long term cooling. The experiment wasterminated one minute after the LPIS injection.

The chronology of various events for Experiment LP-SB-3 is presentedin Table 1. For details of the experiment description, Reference I shouldbe referred.

2.2 Discussion of the Experimental Results

Experimental results have already been presented in the QLR of thisexperiment. In order to perform the post-test analysis and compare withthe data, additional post test analysis was performed to identify primary-secondary relationship and its implications. A summary of the post-testanalysis emphasizing the new findings will be discussed in the followingsubsections.

The experiment is discussed in there sections covering (a) mass

depletion, (b) core boil-off, and (c) core cooldown and recovery phases.

2.2.1 Mass Depletion Phase

This phase refers to the duration of the transient between the break

4

initiation and primary coolant pumps trip at 1600 s.

Figure I and 2 present the short (150 s) and long (7000 s) term primaryand secondary system pressures. The primary system depressurizationunder the control of subcooled break flow until about 100 s. Thedepressurization rate increased slightly when the pressurizer emptied atabout 70 s. The primary system depressurized to saturation conditions inthe break piping upstream of the break at about 100 s after the breakinitiation and the primary-secondarypressure coupling was established.The core reached saturation at about 200 s. Energy addition to the primarysystem from the decay heat and primary pump heat was mainly removed bythe break flow and by heat transfer to the steam generator secondary side.The heat input to the secondary side was relieved by periodic cycling ofthe main steam control valve (MSCV) during the first 1030 s of thetransient. Figure 3 presents the steam generator secondary pressure withthe opening and closure set points. It appears that instrument noise causedthe valve to open or close exceeding or without actually reaching the valveset point. The valve did not seat 100 % nor did it seat the same closure,yielded different leakage from the steam generator during differentperiods of the experiment. After the last closure of the valve, the steamgenerator pressure showed a similar pressure increase to the pressurerise rates occurred during the earlier closures. After it reached apressure some where between the opening and closure set points, thepressure increase ceased and started dropping. Two different mechanismmight be responsible for this. The first is that the steam valve's fourthclosure position let more steam leakage than the previous closurepositions. And the second is that the heat generation in the primary sidewas not enough to cope with the energy loss from the system, which couldalso be related to the first mechanism. The primary and secondary systemsshowed a continuous pressure drop after the fourth MSCY cycling until thebreak was isolated.

After the saturation coolant condition was established in the break line

at about 100 s, the fluid density at the break continued to decrease due to

voidage of the fluid in the break line was increased. Figure 4 and 5 present

the break line density and mass flow rate. Increased voidage in the

primary coolant system decreased the break flow.

5

Fluid densities in the intact loop hot and cold legs are presented inFigures 6 and 7. As a result of MSCY cycling, increased cooling in thesteam generator tube bundle affected the intact loop cold leg density and

the intact loop hot leg fluid velocity which is shown in Figure 8. Thedifferential pressure across the pump continuously decreased as theprimary coolant voidage increased. Oscillation in the pump differentialpressure started at around 700 s at a vapor fraction at the primary coolantpump inlet of 40 %. The oscillations seen in Figure 9 continued until875 s. Although pump operation caused a homogeneous density in theintact loop cold leg, onset of fluid stratification as indicated by thedensity measurements in the intact loop hot leg (Figure 6) occurredshortly after the abrupt drop in pump differential pressure at 875 s. Thepump operation continued until 2800 Kg mass in the system remained. Theactual remaining inventory was 800 Kg more than the desired 2000 Kg as.specified in Experiment Specification Document (ESD) 4 .

2.2.2 Core Boil-Off Phase

This section discusses the core boil-off phase which refers to theduration of the transient between the primary coolant pump trip at 1600 sand the steam generator feed and bleed operation.

Cessation of the pump operation caused also fluid stratification in theintact loop cold leg as seen from the cold leg density measurements asillustrated in Figure 7. This stratification caused the break plane tobecome uncovered. Thereafter, the break flow was high quality steam.Consequently, the primary system depressurization rate was increased.The other affect of the pump trip was on the pump seal density and massinventory. Since the fluid in the suction and as well as in the dischargepiping collapsed with the pump caostdown, the liquid collected in the loopseal caused immediate density increase in the loop seal as presented inFigure 10. The loop seal remained covered afterwards and preventednatural loop circulation. The secondary system pressure followed theprimary response, and both pressures decreased at 0.0 kPa/s rate.

6

A slow boil-off process was taking place in the reactor vessel governed by(a) decay heat energy addition to the fluid, (b) break mess flow, and (c)pnmary-secondary heat transfer rate. The upper portion of the coreshowed dryout at about 3800 s, when the primary system mass inventoryfell to approximately 1750 Kg. The dryout progressed towards bottom ofthe core. Based on the dryout times indicated by the claddingthermocouples, the rate of core uncovery rate was about 1.8 mmls whenthe dryout level was within the upper 1/3 of the core, and was reduced to1. 1 mm/s when the dryout level was between 2/3 and 1/3 of the coreheight. The boil-off rate decreased as the core become uncovered becauseless energy was transferred into the fluid from the lower portion of thecore below the peak-power region which is 0.68 m above the core inlet.The cladding temperature measurements below 0.13 m elevation indicatedthat the core was not completely uncovered. Figure II presents the fuelcladding temperatures in the center bundle. Figures 12 presents the fuelcladding radial temperature response at the 1. 14 m elevation. Differencesin the heat up rates at (a) different axial elevations and (b) at differentradial locations were observed. The QLR suggested that a small amount ofentrained water carried up into the core slowed the heat up rate at higherelevations. The difference in the heat up rate in the radial direction wasexplained with the difference in the radial power factors. Further analysisof the primary-secondary relationship indicated that the indicateddifferences were more likely as a result of liquid drainDack from the hotleg, generated due to reflux condensation in the steam generator tubes andits re-evaporation in the core. Detailed discussion of the primary and thesecondary relationship is given in the following subsections.

2.2.3 Discussion of Primar -Secondary Relationship DuringExperiment LP-SB-3

Post-test examination of the measured pressures and temperatures in theprimary and secondaru sides and the conditions existing in the inlet andoutlet plena providea that reflux condensation was occurring when theprimary side pressure and/or temperature was higher than the secondarypressure and/or temperature during the boil-off phase. This conclusionwas further supported by the RELAP5/MOD2 post-test calculations andwill be presented in Section 3. The possibility of drainback of the

7

condensed liquid through the hot leg into the core and interpretation ofvarious data providing evidence to this conclusion are discussed in thefollowing two subsections.

2.2.3.1 Reflux Condensation and Heat Transfer to the Secondary SidePressure differential between the primary and secondary systems ispresented in Figure 13. As the core uncovery progressed, due to lesserenergy transfer to the fluid, and at the same time due to continuous energyloss from the break, the primary side pressure dropped below thesecondary side pressure at about 4300 s. The primary side pressureremained below the secondary level until about 4900 s, slightly after thebreak isolation at 4750 s. During this time period the secondarytemperature was higher, as illustrated in Figure 14, than the saturationtemperature of the hot leg. Outside of this period it was believed thatcondensation was taking place in the steam aenerator tubes below thesecondary liquid level, especially close to the tube sheet and forcing theprimary pressure to follow that of the secondary pressure. Thecondensation heat transfer in the steam generator tubes was theadditional heat removal mechanism in the primary side besides the heatremoval via the break flow. After the break was isolated, it became theonly heat removal mechanism in the primary system other than the heatloss to the environment. The sharp change in the slope of the pressuredifferential suggests that the condensation heat transfer wasreestablished just after the break isolation, presumably increased inmagnitude to compensate the loss of the primary side energy outflow viathe break flow. Therefore, the condensation heat transfer became anessential factor in the overall heat balance existing in the primarysystem.

The heat transfer from from the primary to secondary side was estimatedto be 50 to 150 kW in the QLR with the lower bound attained which thepressure differential was negative. On an approximate basis, the actualheat transfer can be excepted between zero to 100 kW with a 50 kW offsetto eliminate the heat from the metal work ( which was not taken intoconsideration in the QLR) and the instrument (which indirectly measures

8

the steam generator level) uncertainties. A heat transfer of probable 50kW is equivalent to a condensation rate of 30 g/s 5 in the primary tubes.

2.2.3.2 iquid Drainback and its Effect on Fuel Rod Temperature Some ofthe condensate that formed in the upside might have probably drained backinto the vessel. The indirect evidences of this drainback outside of the4300-4900 s period were (a) the thermocouples in the steam generatorplena showed that saturation conditions persisted in those regions duringthe entire period except from 4200 s to 5000 s, when superheat wasmeasured (the steam superheat shown in Figure 15 was due primarily tohot wall effects), (b) the upper end box thermocouples which showedtypically 200 K lower temperature in fuel module 4 (which is the nearestto the intact loop hot leg) than in equivalent ones in modules 2 and 6, (c)cladding temperatures in modules 4 were asymmetrically lower than inmodules 2 and 6; by 60 K at 1.14 m, by 50 K at 0.76 m, and approximatelyequal at 0.28 m above the core inlet, and (d) the fuel heat up rate, as seenin Figure 12, considerably decreased after 4900 s with thereestablishment of the reflux condensation, and (e) corner module liquidlevel probes, as illustrated in Figure 16, showed that the probe at 1.77 melevation above the core inlet was wet outside of the negative pressuredifferential period (4300 to 4900 s) although the dryout level was well inprogress in the core.

2.2.4 Core Cooldown and Recovery Phase

This phase refers to the duration of the transient starting with theinitiation of the feed and bleed operation at 5415 s and with experimenttermination at 6845 s.

The feed and bleed operation was initiated at 5415 s when the peakcladding temperature in the core reached 967 K. The feed and bleedoperation was affective as anticipated to depressurize the system at anincreased rate and to terminate the cladding heatup and even caused thelower section (below 0.36 m elevation) of the core quenched as seen inFigure 11. The primary system pressure dropped to the accumulatorinjection set point in about 140 s. The fuel cladding cooling induced by the

9

steam generator feed ari bleed operntion was ennfnnced by thre Initiation ofthe accumulator. The core was completely quenched and refilled in 242 safter the accumulator injection started.

3. RELAP5/MOD2 COMPUTER CODE SIMULATION OF EXPERIMENT LP-SB-3

The RELAP5/MOD2 computer code was used for the post-testcalculation of Experiment LP-SO-3. RELAP5/MOD2 is an advanced, bestestimate computer program developed at the Idaho National EngineeringLaboratory (INEL) for the analysis of Loss-of-Coolant Accident (LOCA) andother PWR transients. The specific application of the code to theexperiment LP-SO-3 post-test simulation is discussed in this section.

3.1 RELAP5/MOD2 Description

RELAP5/MOD2 employs a finite-difference fluid cell representation ofthe primary and secondary coolant systems. The six-equationhydrodynamic formulation employs separate equations to describe theconservation of mass, momentum and energy for liquid and steam withineach fluid cell. The description of the hydrodynamics is essentially onedimensional within each fluid cell. The inclusion of a simplified treatmentof the conservation of momentum in the direction perpendicular to themain stream flow, where cross flow occurs between parallel volumes andin branches, brings a special treatment of two-dimensional effects.

Description of the hydrodynamics of choked flow, stratified flow, andabrupt area changes are carried out with special process models. Specialmodels are included for simulation of particular components, such aspumps and accumulators. Flow-regime-dependent constitutive equationsand heat transfer packages are incorporated to complement thehydrodynamic description. Conduction of heat within metalwork or fuelrods is calculated with one-dimensional (two-dimensional in fuel claddingfor reflooding simulation) finite difference formulation. A powerfulcontrol and trip logic capability is built into the code.

10

3.2 RELAP5/MOD2 Nodalization and Input Model for Experiment LP-SB-3

The nodalization used for the base case calculation was based on theRELAP5/MOD I input deck that was used for the planning and prediction ofthe experiment at the INEL. The deck was first converted to theRELAP5/MOD2 format and fallowing changes were applied:

a- to utilize cross flow junction model at the pressurizer connection tothe hot leg to obtain an improved pressurizer drainage,

b- to correct minor errors in the main steam valve flow area, and insome of the control variables mainly calculating heat loss, massetc,

c- to tune various valve opening and closure timing.

Figure 17 presents the nodalization diagram used for the base casecalculation. Some more updates were done during the steady statecalculations and early phase of the transient in order to:

a-match the by pass leak flows between the downcomer and upperplenum and through the reflood assist piping valve with theexperience,

b-tune drainage timing of the pressurizer by increasing flow losscoefficients of the surge line,

c-tune MSCV leak to match the first MSCV cycling time,a-reorientate the by-pass flow junction connecting the downcomer

upper annulus to upper plenum upper volume to avoid subcooled statebeing calculated in the upper two volumes of the upper plenum. Thefinal connection was between the downcomer upper annulus and theupper plenum nozzle area.

'The final version of the input data listing is supplied in Appendix B.

The input model consists a total of 32 fluid cells for the vessel and100 cells for the remainder of the primary, secondary and ECC systems.Characteristics of the model are as follows

a- a split downcomer upper annulus was used with the cross flowconnections,

b- filler gap was separately modeled,c- core was modeled by a average channel approach and represented by

II

six equal length heat slabs,d- reactor vessel internals and loop piping, pressurizer and steam

generator walls were simulated with heat slabs. Heat losses fromthe piping to the environment was modeled,

e- steam leak through the steam valve was specially modeled,e- one velocity break flow model was used at the break. Discharge

coefficients for the subcooled and saturated flows were 0.93 and0.8 1, respectively.

4. RESULTS OF RELAP5/MOD2 POST-TEST CALCULATION

This section presents the thermal-hydraulic results of ExperimentLP-SB-3 post-test calculation. Prior to performing the post-testcalculation, a steady state calculation was performed to obtain the initial.conditions measured during the experiment. Following the steady statecalculation, transient calculation was started with the trip set pointstaken from the experiment. The following subsections discuss the steadystate and transient calculations.

4.1 Calculation of the Steady State

Using a steady state controller package the simulated LOFT system wasbrought to the required initial conditions. At first, steady state option ofRELAP5/1MOD2 was used to automaticallu terminate the calculation -whenthe required steady state was achieved. However, this option terminatedthe calculation at a time when various key parameters were too far offfrom their steady state values. Therefore, the steady state calculationwas performed with the transient option. The calculation was continueduntil the observed variations of the calculated values of these parametersfrom their desired values were acceptable. The key parameters controlledusing the control variables were the primary system pressure, pressurizerlevel, cold leg temperature, primary system mass flow rate and steamgenerator secondary level. The behaviors of the secondary side feed andsteam flows, pump speed and head, pressurizer heater power, pressurizerspray valve and steam generator main steam valve positions, and primaryside charge or let down flows were the other parameters checked for the

12

steady state.

The system pressure was controlled by the pressurizer spray whichinjected cold leg fluid to the pressurizer to reduce the pressure if thepressure was calculated to be greater than the measured value. The secondcontroller on the system pressure was the pressurizer heaters. Theseheaters were charged if the pressure was lower than the set point. Theseheaters, although in realty ware located close to the bottom of thepressurizer, were placed at the mixture level in the RELAP5 model toincrease the boiling. The pressurizer level was controlled by twocontrollers. One controller which charged fluid at the cold leg temperatureto the cold leg if the pressurizer level was lower than the set point. Thesecond controller dumped the system fluid to a time dependent volume ifthe pressurizer liquid level was higher than the set point. The final valuesof the primary pressure and pressurizer level were calculated to be almostthe same as their measured values. The final valve positions controllingthe pressurizer spray, primary system charge or let down flows were zero.The final pressurizer heater power was zero. The pressurizer surge lineflow was negligible at the end of the steady state calculation.

The primary loop flow was adjusted by using a proportional/integralcontroller based on loop flow error to control pump speed. The steadystate intact loop flow was calculated to be the same as the experimentalequivalent. The pump speed and head were in agreement with the measuredinitial values. The broken loop flow (from the vessel to the cold leg and viathe reflood assist valve to the hot leg and back to the vessel) was smalland based on the leak flow through the reflood assists by-pass valve. Thetotal by-pass leak flow based on the flowy loss coefficients used in theINEL deck was calculated to be about 2% of the total loop flow. This valuecompared with the generally accepted 7% of the loop flow was consideredto be too low.

The cold leg temperature was controlled by the main steam valveposition with a proportional/integral control system. Based on the steamflow rate and heat transfer to the secondary side, the code calculated thesecondary system pressure. Another control logic was used to adjust thefeed flow to control the steam generator required level. This controller

13

was also coupled to the main steam flow. The steam generator level, mainsteam and feed water flows were calculated to be the same as measured.Although the steam and feed water flow rates were correctly calculated,the steam generator secondary side pressure was the only parameter beingcalculated offset by 0.35 Mpa from the measured equivalent.

After about 300 s of calculation the steady state was consideredacceptably stable. Two 150 s long steady state calculations as acontinuation to the previous run were performed:

a-to increase the total by-pass leak flows by decreasing the flow losscoefficients of the two leak junctions,

b- to increase the secondary side pressure.

The system reached steady state condition and become stable duringthe first 150 s long transient and the total by-pass leak flow wascalculated to be 4.5% of the total loop flow. Further trials to increase theby-pass leak to the generally accepted level were not done due to thecomputer cost. The final value was accepted to be adequate. Despite of thetrials done to increase the steam generator secondary pressure, nosuccess was reached. The complex geometry and atypical internalstructure of the steam generator with rather simple nodalization werepossible causes of the problem.

4.2 Transient Calculation

After the steady state was considered stable, various steady statecontrollers associated with the pressurizer heaters, pressurizer spray,primary system mass charger and let down, pump speed, and various valveposition were removed. The trips for various actions were defined basedon the measured data. Table 2 compares the calculated and the measuredsteady state values. The transient calculation was started from time zeroand using the last restart record in the steady state plot-restart file. Thecomplete transient was calculated in three major restarts. Each restartused the recent code version available at the time of the run. These werethe cycles 33, 35, and 36.

14

The course of the calculated simulation is discussed in the samephases that were analyzed in Section 2, namely : (a) mass depletion, (b)core boil-off, and (c) core cooldown and recovery phases. Table I comparesthe timing of various events calculated and measured.

4.2.1 Mass Depletion Phase

This phase begun with the opening of the break, defined as time zero.Reactor scram followed rapidly (measured at 9.2 1 s, calculated at10.746 s) when the hot leg pressure fell to the scram setpoint of 14.19MPa. The steam generator main steam control valve was closed and feedwas isolated upon scram. Primary system coolant pumps were left runninguntil the primary system inventory fell to 2800 Kg from an initialinventory of 5600 Kg.

4.2.1.1 Primary and Secondary System Pressure Behavior Calculatedand measured intact loop hot leg and steam generator pressures during themass depletion phase are shown respectively in-Figure 18 and 19.Primary pressure dropped rapidly under the control of the subcooledblowdown. After the pressurizer level dropped below its indicating range(corresponding to approximately 8 cm of the calculated level at this time),the calculated and measured system pressures showed an increaseddepressurization rate. This trend continued until saturation condition wasreached at 99 s in the break line upstream of the break during theexperiment, but ceased in the calculation about 15 s after the pressurizerbecame empty. This reduction in the calculated depresurization ratecaused the simulated system to reach the end of blowdown condition atabout 200 s. The difference between the depressurization rates measuredbetween 70 and 99 s and calculated between 77 and 200 s was believed tobe associated with (a) poor modeling of the pressurizer surge line and (b)the water remaining below the surge line entrance after the pressurizerlevel showed below its indicating range and effect of flashing of thiswater on the system pressure response. It was doubtful that flow losscoefficients representing the entrance effects and also the bends in theline and the nodalization of the 7 meter line with three fluid cells wereadequately modeled. However, for a long experiment having majorphenomena occurring a couple of thousand seconds later, it was believed

15

that this time shift did not affect the subsequent thermal-hudraulicphenomena. The calculated primary system pressure was controlledthereafter by the secondary side pressure as in the experiment.

After the main steam control valve was closed and the feed water wasisolated, the secondary side pressure reached the cycling pressure of 7.12MPa (measured at 87.5 s, and calculated at 76.5 s). The first cycle wasfollowed by three more cycles, the last one being measured at 1030 s andcalculated at 1175 s. The difference between the measured and calculatedcycle times were associated with the different opening and closure setpoints observed during the measurement. After the last cycling, thecalculated secondary pressure continued to increase further, although themeasured pressurize increase ceased shortly after the valve was(supposedly ) closed. This difference was believed to be either as a resultof improper valve seat position which resulted in more leak thencalculated or more energy transfer calculated to the secondary side.Calculated primary and secondary pressures were overall in agreementwith the date except with small deviations due to main steam valvecycling behavior.

4.2.1.2 Break Flow Calculated and measured break flows are shown inFigure 20. The break flow rose from from zero rapidly to a peak of 6.4 Kg/sin the calculation and to 6.7 Kg/s in the experiment than fell to 2.5 Kg/sat the end of the subcooled break flow period and to 1.3 Kg/s just prior tothe pump trip. The break flow calculation except with minor differencesassociated with the pressure behavior was in excellent agreement withthe data.

4.2.1.3 Densitg Comparison of the measured and calculated densitiesin the intact loop hot and cold legs are shown respectively in Figure 21and 22. Similar comparison in the break piping is shown in Figure 23.Overprediction of the density in the hot leg was a result of the differencebetween the calculated and measured secondary system pressures afterthe fourth main steam valve cycle and its consequent effect on theprimary system pressure. The calculated beak line density wassurprisingly well in agreement with the data.

16

3.6.1.4 Pump behavior Calculated and measured pump pressuredifferential are displayed in Figure 24. After the subcooled blowdownended the pump pressure differential showed a continuous decrease. Thecode calculated the pump degradation indicated in Section 2 slightlylater.The oscillations seen in the data were also observed in thecalculated pump performance but being small in magnitude. The pumphomologous curves (mainly were developed based on the Semiscale pumps6-nd partially modified using LOFT L3-6 experimental data 7 ) used for thiscalculation was responsible for this difference. An excellent agreementbetween the calculated pump differential pressure and the data existedafter about 1300 s in the transient.

4.2.1.5 5ystem Mass Inventory Figure 25 presents a comparison of theprimary system mass inventories. An excellent agreement was reached asa result of a good break mass flow calculation. The 2800 Kg Inventory wascalculated at 1618 s and measured at 1600 s.

4.2.2 Core Boil-Off Phase

This phase starts with the pump trip and continues until the peakcladding temperature reaches 987 K to Initiate feed and bleed operation.

4.2.2. 1 Primary and Secondanri System Pressure Behavior Comparisonsof calculated and measured pnmary and secondary pressures are presentedin Figures 26 and 27, respectively. The primary side was depressurizingas the break discharged almost pure steam. The data (Figure 13) showedthat the primary side pressure fell below the secondary level only afterthe core uncovery was started. The code calculated similar trend exceptbetween 2300 and 3000 s. when the secondary side pressure was higherthan the primary side as illustrated in Figure 28. This discrepancy was.related to a trip logic which did not activate the steam valve leakalthough the main steam valve was close. The leak was designed to beactivated whenever the steam generator pressure fell below the mainsteam valve closure set point. After the last cycle, in absence of thisleek, it took quite a long time until the secondary pressure fell below thevalve closure set point end then the leak was activated.

17

The primary pressure was not calculated to fall below the secondarypressure after the heat up started. This was due to late core uncoverycalculation, which caused the simulated primary system received slightlymore heat from the core. The core begun to uncover (measured at 3800 s,calculated at 4008 s) and hence the heat transfer to primary fluid wasreduced. This increased the depressurization rate and the hot legsaturation temperature fell below the secondary temperature (measured at4300 s, calculated at 4325 s) as presented in Figure 29 and 14 where thecalculated and measured primary secondary temperature differences arepresented. The sign of the calculated temperature difference reversed, asin the experiment, shortly after the break isolation at about 4905 s. Thecalculated primary secondary pressure difference was much smaller thanthe measured difference. As a result of this, the primary secondarytemperature difference was less than the measured difference. Figure 30presents the condensation rate (negative vapor generation rate) in thesteam generator upright (Volume 115020000). Figure 31 presents theliquid flow flowing into the steam generator inlet plenum. The codecalculated condensation in the steam generator tubes. The amount of totalcondensate was generated at an approximately 40 g/s rate. A reasonablecondensation axial distribution was calculated in the steam generatortubes, with maximum condensation rate being calculated in the cells nextto the tube sheet and no condensation being calculated in the cells ofwhich outside pipe surface was uncovered. The condensation ceased in thesteam generator tubes except in the very lower cell at about the initiationof the core uncoveru. The amount of condensation calculated might besmaller than what actually occurred during the experiment due to thecalculation of smaller primary-secondary temperature difference.Figure 32 presents the steam and liquid flow rates in the intact loop hotleg (Volume 100010000). The code also calculated condensation in the hotleg due to the heat transfer to the environment. For a very long transient,the piping inner wall should be at the saturation temperature of thesystem couple of thousand seconds in the transient and act as a heat sink.Since there was no data taken to indicate the transient heat loss from thepiping, it was not possible to verify the calculated condensation rate. Theother important conclusion from Figure 32 is that the liquid and steamflow rates were the same indicating that the heat and mass transfer and

18

resultant mass balance were calculated properly. The calculated ref luxcondensation in the steam generator upright was restored after the breakisolation. Although the code was not able to calculate the primarysecondary pressure crossover due to different core uncovery timing, theprimary and secondary pressures and continuation of reflux condensationwere satisfactorily calculated

4.2.2.2 Density Figures 33 and 34 present a comparison of themeasured and calculated densities in the intact loop hot and cold legs,respectively. After the pump trip the liquid in the system fell to the lowerpart of the loop, e.i, to the reactor pressure vessel and loop seals,therefore, a sharp reductions in the densities in the hot and cold leg wascalculated similar to the measured reductions. The code calculated theliquid build up in the loop seal, as illustrated in Figure 35, and the reactorvessel, as illustrated in Figure 36 where the collapsed liquid level in thereactor vessel is shown. The recoveries seen in the measured hot and coldleg densities immediately following the indicated drops were very wellcalculated by the code.

4.2.2.3 Break Flow As noted above, the break flow was of single phasethroughout the boil-off phase, and was very accurately calculated, aspresented in Figure 37.

4.2.2.4 Core Dryout and Thermal Response Figure 38, adopted from theOLR, presents the measured core dryout information based on thethermocouples and liquid level probes and also overlaid are the RELAP coreheat slab nodalization and calculated dryout times. RELAP5/MOD2calculated the dryout initiation in the core approximately 200 s later thenearliest measured dryout based on the thermocouple readings. This figureshows that the dryout time at 1.397 m (55 in) elevation (whichcorresponds to the bottom of the top most heat slab) occurred sometimebetween 3830 s (implied by fuel cladding thermocouple reading) and 41032s (implied by the corner module level probe) (possibly later as will beexplained below). The center bundle thermocouples showed dryout earlierthan the dryout indicated by the center bundle liquid level probes and muchearlier than the dryout indicated by the corner bundle liquid probes.

19

The code (due to the heat transfer logic) waited until the void fractionin a volume associated with a heat slab reached a critical void fraction(which is 0.99999 ) after which it could pass to post-dryout heat transfermode. Therefore, the calculated dryout times were more relevant to thedryout indicated by the liquid level probes than the dryout times of thecenter bundle cladding thermocouples. The late dryout indicated by thecorner channel liquid probe was believed to be a result of the liquiddrainback from the hot leg before 4300 s or after 4900 s. Since this cornerprobe was located under the broken loop hot leg, it could be arguable thatthe condensed liquid draining back from the intact loop hot leg was moreeffective on the dryout of the fuel moduls 7,4 and 8 which were justunder this leg. Since the code considered homogeneous distribution ofthe drained liquid in the nozzle area, which was believed not to be thecase, and homogenous downward liquid flow, the net effect of this drainedliquid postponed the dryout of the core. There were three more uncertainareas which might be responsiole for this late dryout. These are: (a) notaccurate mass distribution in the primary system, (b) over estimation ofthe amount of condensation occurring in the hot leg, and (c) smallerpressure differential between the hot and cold legs.

Since the code predicted a correct mass inventory in the system asseen in Figure 25, the distribution of the mass in the system might bewrong due to artificial liquid hold uo in the hot leg before the dryout begunand afterwards excessive liquid draining back to the core prevented thedryout. Because of the center of gravity of the simulated hot or cold legsare at a different elevation than the center of gravity of the connectedvolume in the vessel, liquid hold up might be a problem due to improperconnection simulation. A sensitivity analysis in which the cross flowconnection of the legs to the vessel was planned to investigate the effectof the cross flow junction model on the dryout initiation.

Since there was no data available to indicate the amount of heat lossfrom the hot leg which determines the amount of condensation, no furtheranalysis was planned.

These junctions might have created a lower back pressure in the upperplenum than might have consequently caused the liquid to be hanging inthe core longer and resulted in late dryout. Since the flow loss

20

coefficients of the by-pass junctions were decreased during the earlyphase of the calculation to increase the by-pass leak flow, anothersensitivity was not planned.

Although the code calculated slightly late initiation of the drjout andslightly higher core uncover'J rate, initiation end progress of the drijoutwas successfully calculated. The calculation indicated that the core wasnever completely dryed out, as was in the experiment.

Figures 39 to 44 present a comparison of the thermocouple data withthe calculated temperatures at levels I to 6. Calculated temperatures atadiabatic boundary condition are also overlaid. It can be seen that theinitial heat up was almost adiabatic, both in the calculated (except at thesixth elevation) and measured data. As significant cladding to coolanttemperature difference was established more heat transferred to thesteam and the heat up rate, although being higher in the calculation, fellbelow the adiabatic rate.

The measured heat up rate at the top of the core was greater than theadiabatic as the steam temperature was above the cladding temperatureand heat was transferred to the fuel. The calculated cladding temperatureresponse at the highest core elevation was different than the measuredresponse due to the followings:

a- the draining liquid was avoiding the heat up of this slab until thereflux condensation in the steam generator upright ceased at about

4500 s. After this time the heat up rate was approximately the sameas the measurement,

b- the reflux condensation in the steam generator uoright started againat about 4930 s in the calculation and caused on affective cooling atthis elevation.

Since the incoming liquid was considerably evaporated in the sixth corevolume, the liquid draining back to the lower core elevations did notaffect the other heat slabs. However, in the experiment, the draining liquidwas probably more effective on the fuel rods in the fuel modules 7,4 and 6after 4900 s. It reduced the cladding and the steam temperatures andcreated considerable radial temperature differential between the center

21

of the core and the filler blocks. Since the LOFT core is a small core wherecenter to filler block distance only 23 times the fuel pitch, thistemperature differential considerably increased the radiation (especiallyabove 700 K) and convective heat transfers especially from the rods inthe center bundle, which in turn reduced the heatup rate extensively asseen in Figures 39 to 41 Since three dimensional behavior of the liquiddrainback into the core was not modeled and the code did not have theradiation heat transfer model, the calculated temperature excursion ratesat the second to the fifth core elevations were much higher than themeasured rate. Therefore, the time at whicn the maximum coretemperature of 816 K was reached (to initiate the closure of the breakvalve) was calculated only 163 s later than the measured 4742 s, althoughthe dryout was calculated approximately 200 s later at the hottemperature elevation.

4*2.3 Core Cooling and Recovery. Phase

At a peak core cladding temperature of 987 K, the main steam valvewas opened to blow the steam generator and at the same time the feedwater flow was restored. This commenced the feed and bleed phase of theexperiment, the time was 5415 s in the experiment and 5166 s in thecalculation. The feed and bleed operation caused the system todepressurize to the accumulator injection set point at 5558 s in theexperiment and at 5360 s in the calculation. Accumulator injectionenhanced the cooling of the system and the system pressure was furtherreduced to LPIS injection set point at 6785 s in the experiment and at6222 s in the calculation. The experiment was terminated at 6845 s whilethe calculations was stopped at 6560 s.

4.2.3.1 Primary and Secondary System Pressure Behavior Feed andbleed in the steam generator caused a rapid drop in the secondary sidepressure, followed closely by the highly voided primary. system, showingthat secondary feed and bleed operation was an affective recoveryprocedure from a small break LOCA with no HPIS injection. Figures 45 and46 present a comparison of the measured calculated primary andsecondary system pressures, respectively The rate of depressurizationwas very well predicted, except with a slight time shifting caused by the

22

calculated core thermal response during the boil-off phase.

4.2.3.2 Core Thermal Behavior With the feed and bleed operation, theincreased condensation in the steem generator caused an increase in thesteam velocity, and therefore entrainment in the core. This, coupled vfit.",the level swell caused by the reduction in the pressure resulted in apartial rewet of the core from bottom up ( Figure 40). The rewet at thecore level below 0.38 m above the core inlet was followed by a re-dryout.This re-dryout seen was not calculated. The upper section of the coreabove 0.38 m above the core inlet started to decrease in temperature. Thecode properly calculated the increase in the steam velocity in the coreThis was evident from the increase in the temperature of the upper mostheat slab (Figure 44). But this increase ceased and temperature of thisheat slab decreased with the increased cooling in the core. The other coreelevations showed a sharp decrease in temperature until accumulatorinjection. The indicated temperature decrease was further enhanced by theaccumulator injection as in the experiment. The difference between thecalculated and measured temperature reduction rates during the periodbetween the feed and bleed operation and accumulator injection was dueto:

a- entrainment calculation,b- steam condensation calculation in the steam generator

upright and its drain back to the core,c- calculated heat transfer coefficient.

The code might have possibly calculated more entrainment whichincreased the cooling excessively at elevation higher then 0.38 m. In theexperiment, possibly most of the entrained water evaporated beforereaching the higher elevations for about 200 s after the initiation of thefeed and bleed. This was then followed by a higher entrainment whichcaused a very sharp reduction in the temperatures.

The code might have calculated more steam to be condensed in thesteam generator tubes and drained back tothe core.

The code might have over predicted the heat transfer coefficient when

the flow contained less than approximately 20 % liquid void fraction

A split core channel approach might have improved the core behavior,provided with a detailed upper plenum nodalization, especially in thenozzle area which should simulate the 3-dimensional licuid drainbackbehavior. However, the difficulty in detailed modeling the LOFT upperplenum which required precise information about the geometry, flow losscoefficients (especially in the radial direction) did not encourage asensitivity study on this issue.

After the accumulator injection started, the core was completelyquenched as in the experiment. The quench progress and temperature, andrate of temperature drop were well in acreement with the data.

4.2.3.3 Accumulator and LPIS Injection The calculated feed and bleedoperation successfully brought the system pressure, first to theaccumulator injection point, ana shortly 2fter to the LPIS injection point(Figures 45) as in the experiment. The calculation went smoothly duringboth injection periods and no noticeably oscillations were observed.

4.2.3.4 Accumulator and LPIS Flow Figure 47 overlays the calculatedand the measured ECC flows. Figure 48 overlays the integrated ECC flows,with a time shift applied to the RELAP calculated flow for a bettercomparison. As seen from these figures, except very small underpredictionduring the first 500 s of the accumulator injection, the calculatedaccumulator flow was in agreement with the data. The slightoverprediction of the LPIS flow was observed.

,. SENSITIVITY CALCULATION

A sensitivity study was performed to assess hot and cold legconnections to the vessel using cross flow junctions. At the time when thebase case calculation was initiated, the original connections of these legsto the veEsel were kept the same since the standard PELAP5 LOFT deck hasbeen extensively used using "down connection" of the legs. This connectionscheme created an artificial elevation difference between the center ofthe gravities of the volumes representing the legs tnd those of the

24

* volumes to which these legs were connected. This unohysical boundarycondition used by the code might have caused a different hydraulicbehavior during the period when liquid from the hot legs was slowlydraining into the vessel. The cross flow junction model was developedespecially for the flow paths where there are no gravity effects involved.Since, a late core dryout behavior was calculated, remodeling of the legconnections using cross flow junctions was tried to see whether anyimprovement could be gained.

In order to use the cross flow junction connections, the upperdowncomer annulus, and upper plenum volumes were remodeled. The finalnodalization diagram is presented in Figure 49. The base case calculationwas repeated with the new nodalization from 3200 s to 5000 s.

The following subsections compare various system parameters

calculated using two different types of the lea connections.

5.1 Pressures

The system pressures in the primary and secondary systems werecalculated to be the same as the ones presented earlier. This was expectedand also was a good indication that the model was functioning properly.

5.2 Core Uncovery

Figure 50 and 51 compare the calculated fuel cladding temperatures atthe fifth and third core levels. The dryout was calculated earlier by about40 s with the cross flow junction connected legs. Since the calculationwas repeated from 3200 s, another approximately 40 s could be gained ifthe calculation would have been repeated from the time of the pump trip.This argument was based on the fact that cross flow junction connectionshould have to have effect during the penod when the not legs and thesteam generator tubes contained considerable amount liquid inventory. Andit should have negligible effect during the time when very small amounrt.ofcondensed liquid was present in the hot leg. This was verified with thefact that there was no difference observed between the calculated heat uprates as seen in Figure 50 and 51.

25

In Section 4.2.2.4, it was indicated that the code predicted the dryoutlate by about 200 s. Application of the cross flow junction model at thehot and the cold legs to vessel connections from the beginning of thetransient would bring this time to approximately to 100 s. This remainingtime difference was believed due to the decreased flow loss coefficientsof the junctions representing the two leak paths which created lesspressure differential between the downcomer and the upper plenum. Anevidence to this conclusion is that as the level progressed into the core,the pressure differential got smaller and the calculated and the measureddryout times got closer to each other.

6. RUN STATISTICS

CPU time versus transient time was shown in Figure 52. The massdepletion phase and early core uncover phases consumed considerableamount of CPU time. The total CPU time was 7.5 hr on CYBER 176 systemfor a total of 6560 s of problem time.. Wherever Courant limit dominatedthe time step size (such as Volume 107), an attempt was made torenodalize that particular volume during the next restart. This problemmainly occurred when the flow in th6 hot leg started to become stratified.The code was forced to run faster by assigning relatively larger time stepsize as the maximum time step size. This caused the code to violate theCourant limit for most of the time. The criteria in selection of themaximum and minimum time step sizes were that number of repeatedtime steps was under a reasonable value. During the base case analysis,10389 time steps were repeated out of a total 114065 time steps. Thegrind timea was achieved to be. 1.7931 ms. The run time of the sensitivitystudy was practically the same as the base case calculation.

a The grind time is the total CPU time diveded by the total number ofvolumes and total number of time steps. A total of 26990 s of CPUseconds was used for 132 volumes and 1 14065 time steps.

26

7. CONCLUSIONS and RECOMMENDATIONS

This LOFT-OECD LP-SB-3 analyses showed that RELAP5/MOD2 correctlycalculated many major system variables in a small break LOCA, such asprmary and secondary system pressures, break flow, density and timingand progression of the dryout. Reflux condensation was calculated in thesteam generator tubes and in the intact loop hot leg.

The major problem encountered was the core heat thermal behavior. Theinitial heatup rate was properly calculated except at the highest coreelevations. The deviations started above 600-700 K at the core elevationsbetween 2 and 5. The heatup rate at the six elevation was more inagreement with the data after the reflux condensation ceased. Although,calculated peak cladding temperature was in agreement, the timing wasdifferent. This agreement was due to the termination of the temperatureexcursion with feed and bleed initiation based on the peak claddingtemperature. The heat up rate in the experiment was controlled by a strongcooling which was believed as a result of the draining of the condensedliquid into the vessel. The condensation was evidently taking place in thesteam generator. Some of the condensed liquid generated in the uprightdrained into the vessel. The resultant downword liquid flow was notuniform across the vessel. Since the LOFT was not instrumented to providedetailed information on reflux condensation occurring in the steamgenerator tubes and possibly in the hot legs, it was not possibleto assessthe code models such as interfacial ,heat'and mass transfer and surfacetransfer, controlling the reflux condensation and itM reevaporation onroute to core. The other constraint was that the non homogeneous liquiddistribution in the upper plenum and its asymmetrical fall back into thecore could'not properly be modeled with a one dimensional code, even withseries of cross flow junction connected volumes w•hich simulate upperplenum and core in detail. The lack of of radiation heat transfer modelincreased the overprediction of the heat up rate in addition to the lack ofproper cooling due to the drained liquid. Therefore, it should not be •expected any better results from the code on the core thermal response.The code properly calculated the cooling induced by the feed and bleedoperation but the amount of it was overpredicted between the third andfifth elevations. Although the partial quench seen in the experiment below0.38 m elevation was reproduced, the code extracted all the stored heatfrom the fuel, and could not calculate the re-dryout at this elevation. The

core behavior (quench progression, temperatures) during the accumulatorinjection period was very good agreement with the measured data.

The pump degradation was properly modeled. The differences observedbetween the calculated and the measured pressure differential across thepumps were as a result of the pump homologous curves used. These curveswere basically developed using the Semiscale pumps and were partiallumodified using LOFT Small Break Experiment L3-6 data.

The problem seen in the calculated heatup behavior associated withimproper simulation of the cooling due to the drained liquid and lack of theradiation heat transfer calculation is believed to be more pronounced inthe small LOFT core. For a large PWR simulation, it is believed that, theRELAP5/MOD2 code will provide a better simulation of a small break LOCAoccurring under similar conditions, provided that an appropriaterepresentation of the physical systems should be provided.

In general, the RELAP5/MOD2 code is adequate to predict the transientthermal-hydraulic behavior of a reactor system undergoing a small breakLOCA. However, the code performance was not only depending on the itsmodels capabilities, but equally depending on how the physical system wasmodeled. The representation of the complicated systems where complexphenomena are occurrng has always given the problem of in what detaileda particular component must be modeled. Attempts to represent everydetail may even create problems due to lack of accurate information suchas geometry, flow loss coefficients etc. Experience developed by using thesame code and the same system for different simulations and assessingwith the experimental data is probably the only guidance on this issue.

8. REFERENCES

I. W. B. Grush et al., _uick Look Report on Oecd LOFT Experiment LP-SB-3OECD LOFT-T-3604, March 1984

2. V. H. Ransom et al., RELAP5/MOD2 Code Manual, EGG-SAAM-6377, April1984.

3. W. B. Grush et al., Best Estimate Prediction for OECD LOFT Small BreakExperiment LP-SB-3, OECD-LOFT-T-3603. February 1984

4. M. Tanaka, M. Peters, OECD LOFT Experiment Specification Document

28

Cold Leg Small Break Experiment LP-SB-3, OECD-LOFT-T-3602,

December 1983.5. Private communication with J. Birchley who was UK delegate to the

LOFT-OECD program, May 1985.6. D. L. Reeder, LOFT. eitm and Test Description, NUREG/CR-0247,

TREE- 1208, July 1978.7. P. D. Bayless and J. M. Carpenter, Experiment Data Reoort for LOFT

Nuclear Small Break Experiment L3-6 and Severe Core TransientExperiment LS- I NUREG/CR- 1868,EGG-2075, Januarq 196 1.

29

TABLE I CHRONOLOGY OF EVENTS FOR EXPERIMENT LB-SB-3

Time AfterExperiment Initiation

(s)

Measured DataEvent

Small break valve openedReactor scrammed on low pressure in intacthot legMain feedwater pump shut off on scram signalMSCV started to closePressurizer level below indicating rangeFirst time MSCV cycled openSubcooled blowdown endedPump differential pressure degradation observedLast time MSCV cycled openPrimary coolant pumps trippedBreak uncoveredStart of core heatupBreak isolatedSteam generator feed and bleed initiatedMaximum cladding temperature reachedAccumulator injection initiatedCore quenchedLow pressure injection initiatedExperiment terminated

0.09.21

9.419.567.87.598.5

700-875.1030.1600.1612.3800. ±_504742.5415.5422.5558.5800.6785.6845.

Post-TestCalculation

0.010.746

10.74610.74694.75.6

201.600-1250.

1155.1618.1630.4008.4905.5186.5198.5360.5570.6222.6560.

30

TABLE 2. 1NITIAL CONDITIONS FOR EXPERIMENT LP-SB-3

Parameter Measured value RELAP-steady state

Primary Coolant System

Power level (MW)Primary coolant mass flowrate (kg/s)Pressurizer pressure (MPa)Pressurizer liquid level (m)Cold leg temperature (K)Hot leg temperature

50.3 ±_1.2

482.6±t2.615.26_0. I1.1 15±0.06556.6±1576.6i 1

50.3

482.615.24

1.094556.6576.12

Secondary Coolant System

Steam generator liquidlevela (m)Pressure (Mpa)Mass flow rate (kg/s)

0.215.58 ±t0.0626.67*_0.77

0.215.21

27.05

.4

a. The level is defined as 0.0 at 2.95 m above the top of the tube sheet

31

is

-Feed Waterisolated

- -Pressurizer empty

N3•

First MSCV cycled open

Su•.• bcooled blowdown ended

as so 75 1"e Its 1s7 9

T I ae Is

Figure 1. Primary and secondary system pressures (0 - 150 s)

to

a rump aegraaatoIUn - - - -

--. Break uncovered

Pumps tripped CO

.Break isolated

tn

s- - SG feed and bleed initia

aAccumulator injection init

6 lose no" 346 40O8

Time (s)

54" 94""S

Figure 2. Primary and Secondary system pressures (0 - 7000 s)

7. S"

- 7. "a

L.2:

W 5

in *@2

0 see 1"S is. U..

Time (s)

Figure 3. Steam generator secondary pressure and MSCV cycling set points

to"

-Sa

o4"*

269

DE-PC-SO5B

Pumps tripped

Break isolated

0

104* z,4 3.04 4a So". I... 7To"0

Time (s)

Figure 4. Fluid density in the break piping

I0."

?. 6s

W0C-

LL.

S *."

a. *so

* 0.

-2.6 0

0 8000 3,g00 so"0 O".

Time (s)

Figure 5. Break flow.

110s

- - - Lower beam, U.-P-I,-UU1a -

I[ ~ -Upper beam, DE-PC-002C

-IPumps tripped A- | I i lt ,

"L, 4"

IAccumulator- ~ -'I 'il"J '' - kill, in,'ectlonl 1,

- I '11 1."' J11

11

•- ~ ~ Jil I ii,

Feed and blee

0 1100 2O 3144 4640o 50 6T0i 7(0s

Time (s)

Figure 6. Fluid density in the intact loop hot leg

s0o"

to - - - Upper beam, DE-PC-OI

coo Pumps tripped

I Accumulator"O 400

Winjection

' " Feed and bleed,

0 1s alsol 3010 480SD S•@t go"q .1

Time (s)

Figure 7. Fluid density in the intact loop cold leg

Is

" .I ,Pump degradation• • .. Fourth time F

FIMSCV opened

Tine (s)

Figure 8. Fluid velocity in the intact loop hot leg

IR*

a"*

' 400

Pump degradation

9-4-)

Ca

QIj

T •

Time (s)

28~

Figure 9. Pressure rise across the primary coolant pumps

1044

Steam generator feed and bleed

400b

0 if" 0 340 40o o 6oo 7Q4

Time (s)

Figure 10. Fluid density in the intact loop seal

1i66

TE-51107-015

5.- -

- T E - 5 G 0 6 -O 1 1 / ' -" " I

Ew"46 -- 60"

o i ( s )

F g r 11 . Fuel c,0t eb

Time (s)

Figure 11. Fuel cladding teniperaturesin the center bundle

1600

TE-4G24-045 -

TE-2G14-045

700

65 4/64"6"94

Time (s)

Figure 12. Cladding temperatures at different radial positions

- * 900

C_

off

Ln

Tie s

Figure 13. Primary-secondary pressure differential

to

4-3

CL

4-atJ

tI-

Ci1

3• 35 4 45f 45f o4soW ss~e

Tic (s)

Figure 14. The temperature difference between the saturation temperature of the

primary system and the temperature of the secondary system

1s

E

0 1000 . 2000 3000 4000 S000 6000 7000

Time (s)

Figure 15. Steam superheat in the steam generator inlet plenuni

a, I I l i I I I 1 1 1 1I1IT - 1 1 1 1III

0 Corner fuel module liquid probe 0 0

Ok~ 0

60 • "Center fuel module liquid probe

• 0 Core dryout

Quenchin~,4A

.0A A 0

S.-0

C-3

as.. 400o 4so 4600 go## 6

Time (s)

Figure 16. Response of reactor vessel liquid level probes, core dryout and quench data

125 Component

R2D Component

525 Sam[5J Pipe with 2 volumesgenerarIs Fn, o°

5Pump

a gsuie Bra~it fnch with junctionsi5- Plassurilef -E

5 5 5I5 415 4 2.2 Juncltion of valve

2 - treatstrlCuctue

-7 Reactor 5

7I3 335 vessel410

S405

3 8 -, 1 m

Figure 17. RELAP5/t1OD2 Nodalization for Experiment LP-SB-3

1~ *oe

52 *

CU0~

10 ~t.

'U1-

~ 3.s*5-

4 * 5@

0 see so". is# e0

Time (s)

Figure 18. Comparison of pressures in the intact loop hot leg

7.60

Lu 'U

di

0 so** "N

Time (s)

Figure 19. Comparison of pressures in the steam generator secondary side

to

a

tn 0

1-~

tS

0dZ:.

6

4

a

S

-a

0 it"o ages to"

Time (s)

Figure 20. Comparison of break flows

1600

-- V Ii u U u eaIMuuL-ri

- - Lo- Upper heamDE-PC

C .. PupstrppdiA

'IA

240 Lower beain,OE-PC-002A

* Pumps tripped-'

I I I I I , i I i , i I I

Tseee 1(s

Time (s)

to

Figure 21. Comparison of densities in the intact loop hot leg

to"1

000

w 40j

0 6" is" is" so"

Time (s)

Figure 22. Comparison of densities in the intact loop cold leg

1046

e oil

all

I3 0

'V.Tim (s)l

sees go"4

Figure 23. Comparison of densities in the break line

-* .. _____PDE-PC-001

4 JC

LnL

Lfl

I ~0 too#I IA IS

Time (s)

IS/S3'131.to"e

Figure 24. Comparison of pressure differentials across pump

I..S

Break isolated>• _• ... _•Dataoo }•D/RELA-

LPIS

z=ee -injection

•- z Pumips

tripped

Accuinu1 at6r

injection

1"4e 2oo 3090. 48H "so So

Time (s)

Figure 25. Comparison of primary system mass inventories

3.00

3.*00

1500 8504 3500 450" mTeI

TIme (s)

Figure 26. Comparison of pressures in the intact loop hot leg

l .06

6.*

L.

4.00

is". 2m0e 44" 55lS

Time (s)

Figure 27. Comparison of pressures in the steam generator secondary side

1. *.

0.O

S-CL

4.00

3.00

15"0 as". 3560 4500

Time (s)Figure 28. Comparison of calculated pressures in the primary and secondary systems

tio. .,

- 1 1O.o -S- - - - -- - - -- -- -

E

a'L--

4~i

40o" 460sees 6eo04o

Time (S)

Figure 29. Calculated temperature difference betwleen tihe primary and secondary systems

1.4

Ul)

E

,k-- -

0b

Volume 11502

' *, 4-0seS so-eI4@44 4SS* $oeo ;see

Time (s)

Figure 30. Vapor generation rate in the steam generator uprigth

6.2O

tj) --

4-)10 4C-IS - - - - - - - - - - - -

0

In

ILLL-I I I-_..LI.I1 I L . L LLLI

3040 30SG 40* 4SO0 S60

Time (s)

Figure 31. Calculated liquid flow rate in the steam generator inlet plenum

6S5"

O.a

11 it4J ~f ~ I/~1~

In

343040666 45606#

TIme (s)

Figure sL.Calculated liquid and vap)or flow rates in the intact loop hot leg

to"

-0 Lower beam, DE.

-- -Upper beam, DE-

see

MEI

4.. ii Pumps tripped

o• 400

02s 3%I0 4 rp.

Time (s)

SS."

Figure33.. Comparison of densities in the intact loop hot leg

too.RELAP5

------- Middle beam, DE-PC-OOIB

Lower beam, DE-PC-OOIA

Upper beam, DE-PC-OOIC

I

EO0

E

ulM

a)

4..

260

,,Pumps tripped

- I 1I f

, , , , , , , , I , , , , , ,I, , , , , , , ,,11 1I i i i,, , , , , , ,

la

-•t*

2566 3N" 454* SIM

Time (s)

Figure 34. Comparison of densities. in the intact loop cold leg

see

E

cri

2SO. 41e" 550 6644

TIne (s)

Figure 35. Comparison of densities ill the loop seal

7.5

s.. -

Top of the core.2:

Buttom; of the core

toee seeso ee ,oo"o

Time (s)

Figure 36. Calculated collapsed liquid level in the reactor vessel

a

cnj

2~e s" 25" 4"6e 66"

Time()

Figure 37. Comparison of break flow rates

?s

q--

-c

w

a,S.0

Li

2s

0

4640 5560 at"0

Time (s)

Figure 38. Comparfson of dryoutSmeasured and calculated

____TE-5106-005

----- ilELAP5(riode-1)

S700

0.0

E

3S"444 4s" 615548)

Time (s)

Figure 39. Comparison of cladding temperatures at 12.7 cm elevation

.-- _____RELAP5 (node-2)- e-- Adiabatic heatup

--------- TE-4t113-015

TE- 5H07-015

E - /--.- -

74 -

354J40" 4600 5S4d 56"J

Tine (s)

0"0

Figure 40. Comparison of fuel cladding temperatures at 38 cm elevation

lose

TE-21115-026S"~/.- "- -,,\-

Adiabatic heatup

S-

S.-

3os" 4 "t 4see SS"N g ,o,"

Time (s)

Figure 41. Comparison of fuel cladding tempera-tures at 69 cm elevation

I*"o

TE-51106-43.8 ----"

-. o--- o Adiabatic heatup /

Ej-7-

6"

/- I , (s)TII- Is

6*O*

Figure 42. Comparison of fuel cladding temperature at 102 cm elevation

Ad iaat/ce /t// p"-/ I,-/

--- Adiabatic heatup

0.o -.0 I/E4 77

600

Time (s)

Figure 43 Comparison of fuel cladding temperat:ures at 114 cm elevation

- 0. I*"

"s

see

4-SLn

E600

460t

3S4 40"0 so" 5500W 60"s

Time (s)

Figure 44. Comparison of fuel cladding temperatures at 157 cm elevation

4. 00

N,\•Feed and bleed

Accumulator injectionNI

14 LPIS injection

" 1 - i LPS i

M%

I I ' I I I , .l I II10 **

ss oo s o 6 oog"o

Time Is)

Figure 45. Ctomparison of pressures in the intact loop hot leg

4.00

CL

tna~LlS1j

S... it.. 655*

Time (s)

Figure 46. Comparison of pressures in the secondary side

is

I.

C)

4-AMu S

0

U-

InL

-I, I I I I i t

see., 550O0 £0.00

Tine (s)

Figure 47. Comparison of ECC injection flow rates

as" To

so4

M_ Data

o - RELAP5I-.-

C0

4-)

CU

*,ma

(T I

Ft Li.

w

ciF SO 80

Ti'Us

Fiue4.Lmaio fitgae C neto lw

5113

Steamgelnerior

125

522

Cormponent

Component

Pipe with 2 volumes

Pump

DOinci with functions

PFesuflzer

515 unction or valve

bliat stniciute

Reactorvessegl r

C)C

350

U

224

Figure 49. RELAP5/MOD2 Nodalization for sensitivity analysis

941

S-p

4-1p

COpCLp

4"0 -2 4446 9 "

Tim (s

Figure 50. Comparison of fuel cladding tempeartures at level 5

------- cross flow juniction connection ror tie legs

m/

- /

4i -co 70/ -

-U -ri

+ee400 44SO 4600 4n9" S"o

,Time (s)

Figure 51. Comparison of fuel cladding tbipeartures at level 3

30444

2zoo#

02

U)

a....

Li

Its" 3"0 5. V ____-..pw .so ,-*

Tie* ($I!)

Figure52 . CPU time versus transient timae and required- CPU seconds per. real second. ..

APPENDIX A

SYSTEM CONFIGURATION

The LOFT facility was designed to simulate the major components andsystem responses of a commercial pressurized water reactor (PWR) duringa LOCA. The experimental assembly includes five major subsystems whichhave been instrumented such that system variables can be measured andrecorded during a LOCA simulation. The subsystems include the reactorvessel, the intact loop, the broken loop, the blowdown suppression system(BST), and the ECC systems. Complete information on the LOFT systems isprovide in Reference A-I and a discussion of the LOFT scaling philosophyis provided in Reference A-2. details of LOFT measurement capability andits instrumentation can be found in Reference A-I.

The arrangement of the major LOFT components is shown in FiguresA-I.The intact loop simulated three loops of a commercial four-loop PWRand contains a steam generator, two primary coolant pumps in parallel, apressurizer, a Venturi flowmeter, and connecting piping. A spool piecewas connected to the intact loop cold leg downstream of the pumpdischarge. A schematic of the break piping and nozzle configuration isshown in Figure A-2. The break piping provided the break path during theblowdown. This line was closed when the maximum temperature in thecore reached 977 K.

The broken loop consists of a hot leg and a cold leg, each of which areconnected to the reactor vessel. The broken loop hot leg also contains asimulated steam generator and a simulated pump. These simulators havehydraulic orifice plate assemblies which have similar (passive)resistances to flow as an active steam generator. Both of the legs havequick opening and isolation valves. During Experiment LP-SB-3, thesevalves remained closed because the break was in the intact loop cold leg.

The two LOFT ECC systems are capable of simulating the emergencyinjection of a commercial PWR. They each consists of an accumulator, ahigh-pressure injection system, and a low pressure injection system.

84

Du'rint Exreriment LP-SB-3, the high pressure injection system was notused. Accumulator 'A" was used to inject a volume scaled amount of !iauidto represent the liquid volume of three accumulators in the referencethree-loop PWR.

The LOFT steam generator, located in the intact loop, is a verticalU-tube design steam generator. The main steam valve is automaticallyregulates the secondary side pressure within given limits.

The reactor contains 1300 unpressurized fuel rods in five square (15 bu15 assemblies) and four triangular (containing 76 fuel rods) corner fuelmodules, shown in Figure A-3 and described in Reference A- 1. The centerfuel module is highly instrumented. Two of the corner and one of thesquare assemblies are not instrumented. The fuel rods have an activelength of 1.67 m and an outside diameter of 10.72 mm.

A- I References

A- i. D. L. Reeder, LOFT System and Test Description, NUREG-CR-02d7.TREE- 1208, July 1978.

A-2. L.J. Ybarando et al., Examiration of LOFT Scaling, 74-WA-hT-53.Proceeding of the Winter Meeting of American Society ofMechanical Engineers, New York, November 1974, CONF-741 104.

Inlact loop Broken loop

" BL. Steam . .. . .. . . :.. Quick opening• .n .generator I - valve (2)experimental simulator

. .. xpe~lmmlla~l Dovwncorner I

mIneasurement . .-- injection -,. " -.s,,,ion. IStGeam location /-"

" It3e ,eat(: a= / vAai-vr :• F t2 ;- ,- " ,'-,..•• • ..

Pressurizer.- . Pu i

- ,. - ..-. . .nm a

0) Drag c•. " . ECC cold 1g n ast• n.en, tand turbint -" \ :., i . injection ,sai.;,n.. .Pumps Ilocation

PC.•3 . .Ye)..erint,-,•,' Reactormelsurement . " vessel .. Suppress:or,

station . \\ ,. k • . vessel Downco.er-Break nozzle " -- ,.

CV.P.139.57 .. "Core - .

\..£ .. Lower pl"numn:

CV-P139.58 V..

Reactor vessel

INEL-L P S1B,3.1510IU DZ)I

Fiqure A-1. LOFT configuration for c'ld 1(. in;!..t loop •,l1i break Li'-SB-3

2.64 cm(1.04 in.) Z

0.4.-4cm R(0.!7 in.)

\0.88 cm

(0.35 in.),

From PC.1, intact3.08 cm(1.21 ir.1

loop cold leg .

1.25-in.dia Ic

Break nozzle insertDetail 'A'

C'j.

Gamma drnsitometer spool pieceDE.PC-S05A. B 8DE.PC-SC5ABD. BD"

To BST

CV-P139-58"C _

• C.P139

.5 7

-TE PC-S09TE-PC-S0OPE-PC-SO10POE-PC.SO5DST Spool piece

FE.PC.S04ME-PC-S04

C.

r.~j PE.PC.SO11

dia

Break nozzle(see Detail A)

I.NEL-LP-Sa-3. 12504

Figure A-2. Experiment break piping instrument configuration.

87

000O

Broken loop - 4

cold log .. ..

t_;_•_.I .;. ....II':" a1II" It

-' . -"- -•

leg 'P111 'J t"12"

L... o.ffiI . .

I': . ,- Vt •'- !' -. ,., , , .. . .•7'i ... .7!.._!;.;!'. ••L~tt~l. '''IL

Inac loop " ' iho .l._e..g --

-'I

4.

-I,

4.

4-

'a-

.1Broken loop

hot leg 1

! ... LCore posi

L. Fuel

B I

Ilion Fuel modulo types

module deslgnation Index

Ii a'!... -'',ii, la:, t

1 j T

a- a1" .....'...'-*4

*li - -:.

a. .. : *ti.'1

..L; _,.-Intact loop

.... ....... leg'-- -* I ' • I *

I

iI - r -b.

f]

0

Fuel fr."i

"Fl,•t'.T.oc ",'ptl 196)h.,,-.Jc(touplC lower tic plate It 7i

[).:nrr . er ttlarT;OcOtlA(.

Na;ut on flux detector. fixed (4)Neluutron flux scan (4l

Ca-iduct'vi:y hqta:d level detector 131.1-iii

ItJEL-A.t6 43?

1:1(ti r I: A- 3. r J.!. ('T ra •,•' n i"•r *f i ti ,,;- ;i ndl Ins;tirnm,'ntt.iti fn

%RC FORM 336- U.t. NUCLIAR REGULATORY COMMISSION I RiPOrT NUMGER (A..,#,Wav riot d WV Ao. " vWy v vi

Q2 841NACM 1102.320M. J2, BIBLIOGRAPHIC DATA SHEET NUREG/IA-0018

sit INSTRUCTIONS ON THE RIVIERS$

2 TITLE AND SUITITLE J LEAVE CIANK

RELAP5/MOD2 Assessment, OECD-LOFT Small BreakExperiment LP-SB-3

4 DATE REPORT COMPLETED

MONTM I TEAR

, AUT.ORISI February 1986AJ OAT(! IEPORT ISSUED

S. Guntay MONC I E ,ARApril 1990

7 PERFORMING ORGANIZATION NAME AND MAI"ING AOODRESS tI.&Z•, Z C* 1 8. P"OJICTITASKIWORK UNIT NUMSER

EIRCH-5303 Wurenlingen 9 IN OR GRANT NUMOIR

Swi tzerl and A6827

10 SPONSORING ORGANIZATION NAME AND MAILING ADDRESS IlmhIw Z Code)I Ili. TYPE OF REPORT

Office of Nuclear Regulatory Research Technical ReportU.S. Nuclear Regulatory Commission 6.,, COV,., .ft.-

Washington, DC 20555

12 SUPPLEMENTARY NOTES

I1. AGSTRACT (9W' A.* be PO

An analysis of the experimental results and post-test calculations usingRELAP5/MOD2 carried out for OECD-LOFT small break experiment LP-SB,3 arepresented. Experiment LP-SB-3 was conducted on March 5, 1984 in theLoss-of-Fluid Test (LOFT) facility located at the Idaho National EngineeringLaboratory (INEL). The experiment simulated a small cold leg break, withconcurrent loss of high pressure injection system, and cooldown and recoveryby feed and bleed of the steam generator secondary side and accumulatorinjection respectively.

This report documents a short post-test analysis of the experiment emphasizingthe results of additional analysis performed during the course of this task.RELAP5/MOD2 input model and results of the post-test calculation aredocumented. Included in the report is the results of a sensitivity analysiswhich show the predicted thermal-hydraulic response to a different input model

IA DOCUMENT ANALYSIS a KSWONIS'OSCRIPTDOS

RELAP5/MOD2, ICAP Program, Small Break LOCA, LOFT

a fO(NTIPI(RsiOPEN| OeO TERMS

UNITED STATESNUCLEAR REGULATORY COMMISSION

WASHINGTON, D.C. 20555

OFFICIAL BUSINESSPENALTY FOR PRIVATE USE. $300

SPECAL FOURTH-CLASSRATEPOSAGE& FESPAID

1 USNRCj